Md. Modabbir Mostofa
Research Subjects
Md. Modabbir Mostofa has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Abu Bakkar Siddique
Maternal and Child Health Division, International Centre for Diarrhoeal Disease Research, Bangladesh (icddr,b), Dhaka, Bangladesh
1 shared publication Ahmed Ehsanur Rahman
1 shared publication- Anindita Saha1 shared publication
Aniqa Tasnim Hossain
1 shared publicationAnisuddin Ahmed
1 shared publicationAnubhab Chakraborty
1 shared publication
Latest Publications
- Open Access
- 4 September 2026
- Published by Elsevier in SSRN Electronic Journal
Performance of InSilicoVA and InterVA-5 for Assigning Causes of Death from Verbal Autopsies in Bangladesh: A Validation Study Using Physician-Coded Verbal Autopsy as the Reference Method
Background: Most deaths in Bangladesh occur in the community without medical certification. Computer-coded verbal autopsy (CCVA) offers a scalable alternative to physician-coded verbal autopsy (PCVA), but InSilicoVA and InterVA-5 have not been compared with PCVA using the same WHO 2016 interviews across all ages. Methods: We did a secondary analysis of 3957 verbal autopsy interviews from urban (n=1948) and rural (n=2009), 2018-2022. PCVA causes, assigned by two independent physicians with adjudication, served as the reference standard. InSilicoVA and InterVA-5 were run using openVA. We assessed CSMF accuracy, individual-level agreement (exact agreement, Gwet’s AC), and cause-specific performance (sensitivity, specificity, PPV, NPV, F1), and reviewed discordant records for misclassification pathways. Findings: Individual-level agreement with PCVA was 47% (InSilicoVA) and 37% (InterVA-5) urban (Gwet's AC 0.45 and 0.35) and 51% and 48% rural (AC 0.49 and 0.46); on this measure InSilicoVA equalled or exceeded InterVA-5 in every stratum at both sites. CSMF accuracy was 65% (InSilicoVA) and 60% (InterVA-5) urban and 75% for both rural, ranging 35-84% urban and 46-71% rural across age groups, with InterVA-5 higher only among rural neonates and under-5. In 18 of 20 discordant cases, supporting evidence lay only in free-text narrative or health worker statements, which neither algorithm reads. Interpretation: InSilicoVA and InterVA-5 are more reliable for population-level surveillance than individual cause assignment in Bangladesh. CCVA can strengthen mortality surveillance, but performance should be validated locally, and targeted VA instrument and algorithm revisions could reduce misclassification.
- 4 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessPerformance of InSilicoVA and InterVA-5 for Assigning Causes of Death from Verbal Autopsies in Bangladesh: A Validation Study Using Physician-Coded Verbal Autopsy as the Reference Method
Background: Most deaths in Bangladesh occur in the community without medical certification. Computer-coded verbal autopsy (CCVA) offers a scalable alternative to physician-coded verbal autopsy (PCVA), but InSilicoVA and InterVA-5 have not been compared with PCVA using the same WHO 2016 interviews across all ages. Methods: We did a secondary analysis of 3957 verbal autopsy interviews from urban (n=1948) and rural (n=2009), 2018-2022. PCVA causes, assigned by two independent physicians with adjudication, served as the reference standard. InSilicoVA and InterVA-5 were run using openVA. We assessed CSMF accuracy, individual-level agreement (exact agreement, Gwet’s AC), and cause-specific performance (sensitivity, specificity, PPV, NPV, F1), and reviewed discordant records for misclassification pathways. Findings: Individual-level agreement with PCVA was 47% (InSilicoVA) and 37% (InterVA-5) urban (Gwet's AC 0.45 and 0.35) and 51% and 48% rural (AC 0.49 and 0.46); on this measure InSilicoVA equalled or exceeded InterVA-5 in every stratum at both sites. CSMF accuracy was 65% (InSilicoVA) and 60% (InterVA-5) urban and 75% for both rural, ranging 35-84% urban and 46-71% rural across age groups, with InterVA-5 higher only among rural neonates and under-5. In 18 of 20 discordant cases, supporting evidence lay only in free-text narrative or health worker statements, which neither algorithm reads. Interpretation: InSilicoVA and InterVA-5 are more reliable for population-level surveillance than individual cause assignment in Bangladesh. CCVA can strengthen mortality surveillance, but performance should be validated locally, and targeted VA instrument and algorithm revisions could reduce misclassification.
- Open Access
- 14 March 2026
- Published by Springer Nature in BMC Infectious Diseases
- Vol. 26 (1)
Barriers and facilitators to dengue prevention strategies in Bangladesh: a scoping review with comparative analysis of other dengue-prone regions
Dengue fever poses a major public health issue in tropical and subtropical areas, with Bangladesh facing a significant rise in dengue cases during the 2023
- 14 March 2026
- Published by Springer Nature in BMC Infectious Diseases
- Vol. 26 (1)
Open AccessBarriers and facilitators to dengue prevention strategies in Bangladesh: a scoping review with comparative analysis of other dengue-prone regions
Dengue fever poses a major public health issue in tropical and subtropical areas, with Bangladesh facing a significant rise in dengue cases during the 2023